GlobalData Head of Thematic Research Cyrus Mewawalla comments, “Companies who invest in the right themes become success stories. Those that miss the big themes end up as failures. These companies are best positioned in the semiconductor sector to profit from the AI theme over the next two to five years.

“AI systems need to process massive amounts of data very quickly. While the performance of general-purpose chips has improved sufficiently to kick-start a new generation of AI technology, they cannot keep up with the dramatic growth in the AI sector, along with the accompanying exponential increase in the volume of data that AI systems process.

“There is a high degree of specialization among available acceleration chip architectures, each focusing on solving performance problems specific to different use cases, deployment environments, and even phases of the AI application lifecycle – modelling, training, inference, optimization, etc.

“This specialization and the popularity of AI itself has led to an arms race among chip manufacturers seeking to solve both narrow and wide AI problems on mobile devices, at the network edge, in the data centre and on the public cloud platform.”

According to GlobalData’s latest Semiconductor Sector Scorecard, AI is one of the top 10 themes that will impact the semiconductor sector over the next two years. The other nine are high performance computing, data centres, ambient commerce, autonomous vehicles, Industrial Internet, gaming, 5G, M&A and geopolitics.

Other semiconductor companies given the highest thematic score for AI are Xilinx, STMicroelectronics, Texas Instruments, AMD, ASML, Baidu, Rohm, Graphcore and Cambricon.